Light enough to forget
Around ≈400 MB in everyday use, so Lithe can stay in the background instead of being closed to free up RAM.
Write Java without paying for the IDE in RAM.
Lithe is a lightweight Java IDE for macOS and Windows. It keeps the parts of IntelliJ IDEA you actually use every day, in about 1/5 of the memory.
order-serviceOrderApplication
OrderService.java HEADLocal changes HEAD
Local changes
Terminalzsh
order-service % mvn test
[INFO] Scanning for projects...
[INFO] Running com.example.order.OrderServiceTest
[INFO] Tests run: 8, Failures: 0, Errors: 0, Skipped: 0
[INFO] BUILD SUCCESS
order-service %
src/main/java›com.example.order›OrderService.javaUTF-84 spacesJava 21
Open a file, search the project, or explore the sidebar.
A Java IDEyou can leaveopen all day.
Browsing code sits at 200–300 MB. A normal working session runs around ≈400 MB, with peaks of 700–800 MB when services are running.
The IDEA core you know
Project structure, editor, search, navigation, and diff. Maven and Gradle projects open as they are.
Ready for AI-written code
When an AI tool edits your files, read the diff in the file it belongs to before you commit.
MEASURED
Three readings from a real workday.
Resident memory on the same Spring service project, taken after indexing completes.
BUILD PRINCIPLES
How we build Lithe.
Two constraints guide every feature before it reaches the workbench.
Every megabyte has to earn its place.
Language services, terminals, build tools, and database helpers start only when the work asks for them.
- BASELINE MEMORY
- 300–400 MB
- TOOLS AND SERVICES
- ON DEMAND
Keep the habits. Remove the weight.
Project navigation, search, Maven, Gradle, run, debug, and side-by-side diff review stay where Java developers expect them.
- CORE WORKFLOW
- IDEA-STYLE
- PLATFORM BUILDS
- MAC + WIN
-
James Whitaker Backend engineer Project tree, navigation, and diff sit where I already look. Switching over barely took any relearning, and the memory drop is obvious.
-
Priya Nair Spring Boot developer IDEA plus Docker used to fill the machine. I can leave Lithe open all day now, and the fans barely spin.
-
Daniel Reeves Software engineer The editor and file tree feel familiar. For day-to-day Spring work I do not need a full IDEA instance sitting in the background.
-
Aimee Park Independent developer Search and jump-to-definition are where I expect them. Walking an unfamiliar service is much easier when the IDE stays out of the way.
-
Marcus Grant Systems engineer After an AI agent touches several files, the side-by-side diff stays next to the code. I can check the result before committing.
-
Amir Shah Full-stack engineer Everyday use sits around 400 MB. Browser, containers, and a local model can finally stay open together.
-
Hannah Cole Platform engineer This is the Java IDE I am fine leaving open all day. It is light, but the daily workflow is still there.
-
Marta Silva Java developer Existing Maven projects open as they are. I spend less time configuring the editor and more time reading the service.
PROJECT SUPPORT
Support that ships with the work.
Partners who keep independent development moving, and the community that gives Lithe a place to grow.
Sponsors
04 active partners 01 
FastAIMODEL ACCESS Multi-model relay access that supports Lithe's AI-assisted development experience. 02 
CodeZGPT ROUTING Stable GPT-family model routing for developers integrating AI into practical tools. 03 
Yuanliu TokenMULTI-MODEL API Flexible relay access to multiple model APIs for experimentation and AI application development. 04 
TorchAIAI INFRASTRUCTURE Model relay services for coding, content, and automation workflows.
FastAIMODEL ACCESS Multi-model relay access that supports Lithe's AI-assisted development experience. 02 CodeZGPT ROUTING Stable GPT-family model routing for developers integrating AI into practical tools. 03 Yuanliu TokenMULTI-MODEL API Flexible relay access to multiple model APIs for experimentation and AI application development. 04 TorchAIAI INFRASTRUCTURE Model relay services for coding, content, and automation workflows. FAQ
A few things worth knowing.
Short answers about Lithe, compatibility, and getting started.
01What is Lithe?
Lithe is a lightweight Java IDE for developers who keep an IDE open all day but do not need a full heavyweight environment. It keeps the IDEA-style workflows you use most — project navigation, editing, search, Maven, run and debug, Git diff — and starts everything else only when you ask for it.
02How much memory does it actually use?
Around 200–300 MB once a typical project is open, and roughly ≈400 MB in normal day-to-day use. Running several services at once pushes the peak to 700–800 MB. Actual usage varies with project size and which services you have running.
03Is it a replacement for IntelliJ IDEA?
For everyday reading, editing, navigating, running, and reviewing Java code, it is meant to be. For deep refactoring, the full inspection set, and the plugin ecosystem, IDEA is still ahead — that is a deliberate trade, and it is what keeps Lithe small. Many developers keep both and reach for IDEA when they need the heavy tooling.
04What does it support?
Java and Spring Boot are the focus, with Maven and Gradle projects opening as they are. Releases bundle the Eclipse JDT Language Server, so there is no separate setup. Entry points are auto-detected for Spring Boot, Java, Maven, Gradle, npm, Cargo, Go, Python, Make, Docker Compose, and shell projects. Java features need JDK 17 or later.
05Which platforms are supported?
macOS 13 or later on Apple silicon and Intel, plus Windows x64. On macOS you can also install and update through Homebrew.
06Is Lithe open source?
Yes. Lithe is developed in the open under the Apache-2.0 license. You can inspect the source, report issues, and contribute on GitHub.
07Where can I report a problem or request a feature?
Open an issue in the Lithe GitHub repository with the steps to reproduce, your OS version, and any useful screenshots or logs.